NUis this AI generated?

Text AI detector

Inspect written text.

Paste at least 60 words and review a writing-style warning signal. Never use this result to infer authorship.

  • Free to use, no account and no sign-up. Heavy automated use is throttled so the checker stays available
  • Documents are split into text sections and embedded images
  • Sampled video submits up to 5 representative frames

Checking API availability…

What this result can tell you.

The detector looks for patterns associated with synthetic or generated content. It cannot prove who made a file, which tool was used or whether the content is truthful.

  • Use the signal to decide what to inspect next
  • Review source and Content Credentials when available
  • Do not accuse a person or block content from this result alone
  • Unknown is a valid outcome
Signals, not proof.

Uploads are temporary and are not used for model training by default. Avoid sensitive, unlawful or children’s media.

Read the privacy draft
Also on your phone and in your browser.

The same detector as an Android app or a Chrome extension — both direct downloads.

Get the apps

AI text detection guide

Review writing-style signals without guessing authorship.

Text detection is especially sensitive to genre, length, language and editing. Use it as triage for further review, never as standalone evidence.

What the text checker does

The checker evaluates a coherent passage for patterns associated with generated writing and returns a warning-only result. It does not search a universal database of AI text, identify a specific model or determine who wrote the passage.

The minimum input protects against meaningless results from headings, short replies and lists. Longer input is not automatically conclusive: a document can mix original writing, quotations, templates and edited generated material.

Prepare a useful sample

  • Paste at least 60 words from one coherent section
  • Keep the original language and punctuation
  • Remove navigation, signatures and repeated boilerplate
  • Analyse distinct authors or sections separately

Do not combine unrelated passages merely to reach the minimum. That creates an average without a clear interpretation.

Why text detector results require caution

Human writing can be concise, predictable or highly polished. Legal clauses, product descriptions, school assignments, translated text and corporate templates may share stylistic regularities. Grammar correction and extensive editing can also change the measurable style. Generated text can meanwhile be rewritten until familiar patterns weaken.

A warning therefore cannot establish misconduct, plagiarism or authorship. If the context matters, review citations, factual consistency, document history, drafts and the writer's ability to explain their process. Give the person a meaningful way to respond.

Responsible uses

Editorial and content review

Use the signal to select passages for source checking, disclosure review and fact verification. Keep the final editorial decision with a reviewer.

Education and hiring

Do not automate sanctions or rejection. Compare drafts and references, assess the work itself and document why additional review was necessary.

See the practical guide to interpreting an AI detector score, the trust and governance principles and the published limitations.

AI text detector FAQ

Can this identify who wrote the text?

No. A style signal cannot establish authorship or prove use of a particular AI tool.

How much text should I paste?

Use at least 60 words from a coherent passage. Several paragraphs usually provide more context than a title or list.

Why might human text be flagged?

Templates, formulaic genres, editing, translation and predictable short passages can resemble generated writing.

Can I use the result for disciplinary action?

Not by itself. Review drafts, sources, revision history and the person's explanation, with an appeal route.

Use the AI text detector Check an image